Rueckl, J. G.
Haskins Laboratories and the University of Connecticut
Studies of novel word learning in adult readers have demonstrated the importance of how novel words are processed during the learning stage, how learning is assessed, and when it is assessed. In some of our recent research we have focused on another factor: the role of the knowledge that that the learner brings to the learning situation. Specifically, we’ve asked whether learning a new word depends on its similarity to familiar words. We’ve operationalized ‘similarity’ at both the lexical and sublexical levels. (Lexical: orthographic and phonological neighborhood size; sublexical: the sequential probability of the constituent letters or phonemes). In our central experiment we contrasted the learning of ‘wordy’ and ‘less wordy’ novel words, intentionally confounding orthographic and phonological similarity to maximize the difference between conditions. Not surprisingly, we found that more wordy novel words were learned faster at the form level. More interestingly, we also found that the meanings associated with more wordy forms were learned faster too. This pattern is readily accommodated by the lexical quality hypothesis (Perfetti & Hart, 2001) and, at a more mechanistic level, the triangle model (Harm & Seidenberg, 2004). Ongoing experiments are aimed towards disentangling orthographic and phonological contributions to this effect and isolating its neural bases.